Advisory Committee on Nuclear Waste 132nd Meeting, February 7, 2002
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: 132nd Advisory Committee on Nuclear Waste
Docket Number: (not applicable)
Location: Rockville, Maryland
Date: Thursday, February 7, 2002
Work Order No.: NRC-213 Pages 1-89
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers
1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005
(202) 234-4433.
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
+ + + + +
ADVISORY COMMITTEE ON NUCLEAR WASTE (ACNW)
+ + + + +
BRIEFING ON
STATUS OF ESTIMATING PERFORMANCE OF IGNEOUS ACTIVITY
+ + + + +
THURSDAY
FEBRUARY 7, 2002
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
The Commission met in at the Nuclear
Regulatory Commission, Two White Flint North, Room
T2B1, 11545 Rockville Pike, at 1:30 p.m., B. John
Garrick, Chairman, presiding.
COMMISSIONERS PRESENT:
B. JOHN GARRICK, Acting Chairman
MILTON N. LEVENSON, Member
. ACNW STAFF PRESENT:
HOWARD J. LARSON, Special Assistant, ACRS/ACNW
RICHARD K. MAJOR
LYNN DEERING
ANDREW C. CAMPBELL
LATIF HAMDAN
SHER BAHADUR, Associate Director, ACRS/ACNW
JOHN T. LARKINS, Executive Director, ACRS/ACNW
AMARJIT SINGH, ACRS Staff
ALSO PRESENT:
BILL HINZE, ACNW Consultant
. P-R-O-C-E-E-D-I-N-G-S
(1:30 p.m.)
CHAIRMAN GARRICK: Come to order. As we
announced this morning, my name is John Garrick,
accompanied by Milt Levenson, another member of the
committee, and Bill Hinze, a former member of the
committee and consultant. We are short a couple of
members of the committee -- George Hornberger, the
Chairman, for compelling reasons, and Ray Wymer, Vice
Chairman, for other reasons -- so we're going to have
to do the best we can. This topic we're going to
start off with, the member who has the lead on it
happens to be Hornberger but, in his absence today,
we're going to lean heavily on former member Bill
Hinze and his distinguished knowledge about this
topic.
So, with that, I think we will go ahead
and start. Tim, are you going to start it?
MR. McCARTIN: And end it hopefully.
(Laughter and simultaneous discussion.)
MR. LARSON: If I were introducing this
topic, I'd say, "This is the cat, the cat that has had
nine lives, or more".
MR. McCARTIN: Thank you, Dr. Garrick.
(Slide)
Today I'll be talking about the status of
estimating performance for igneous activity, and
you're right in that I'm starting it. I will give the
entire presentation, and let me just say the reason
for this is, what we are hoping to do today is give a
perspective not so much in terms of the detailed
geologic igneous processes involved, but more how
we're representing it in the performance assessment
and what it means in terms of its implications in
terms of the dose.
We'll be talking about the differences
between ourselves and DOE, et cetera, but today what
we're hoping to do is talk more about how things are
abstracted in the performance assessment and not
really the detailed geologic processes, why I'm here
rather than someone like Brit Hill from the Center of
John Trapp from the staff. And John Trapp is notable
by his absence. He is recovering from some recent
surgery, which is why he's not here. I'm sure he'd
rather be here than recovering from surgery, that's
for sure.
(Simultaneous discussion.)
MR. McCARTIN: In terms of how I'm going
to present things, it's not so much in order, but the
topics that I'll address, and what I'm hoping to do is
give some small insight on the uncertainties in
estimating volcanic disruption of the repository, talk
to differences between the NRC and DOE in our
approaches for representing this in the performance
assessment, talk to a path forward and, at the very
end -- and I'll do those three topics in the context
of the "big ticket" items for estimating the
performance of the repository.
At the end, I'd like to talk briefly about
the treatment of uncertainty in NRC's TPA code, and
I'm going to draw parallels to what we do in igneous
and in the waste package area. It's the beginning of
an effort to scope out how we're dealing with
uncertainty in the entire code, and we hope to
certainly pull out are there discrepancies. We want
to see, as Marty Virgilio talked about earlier today,
we want to see a consistency in the way we're dealing
with uncertainties, get a better handle on it -- it
might get to some of the treatment of conservatism, et
cetera -- but that's at the very end, and it's a small
nugget. We probably will be coming back to it maybe
later in some subsequent meeting when we have it
fleshed out for the entire code.
(Slide)
First, as just a general overview -- and
this pretty much gets to what I'll be talking about --
what are the areas of uncertainty? Certainly, we have
estimating the probability, and then after you get, if
it will occur, the consequences, and there we have --
I'll talk to five particular topic areas -- the
interaction of magma in the repository, magma in the
waste package, magma in the waste form, the
redistribution of volcanic ash after the event has
occurred, you have an ash deposit, how may it
redistribute in time, and then, ultimately, for the
exposure scenario, the inhalation is the dominant
pathway. And so I'll be talking to each one of these
with respect to those -- the uncertainties, the
approaches, and the path forward.
CHAIRMAN GARRICK: Are there areas that we
should have our attention called to that have gone
through some change in the last few months?
MR. McCARTIN: I'll try to address that.
I think there have been some changes in the DOE
program. I certainly will be talking to things that
we're in the process of examining and I think in the
next six months to a year, we -- I won't guarantee
that we have changes, but we'll have hopefully at
least increased confidence in the approach we have if
it doesn't change. But I will try to talk to that as
I go through.
Right now, in terms of resolutions, there
are agreements in place for all these that we think
the DOE will address the issues. However, one thing I
think the staff certainly recognized that we aren't
going to reduce all the uncertainties in any area of
the TPA code, and uncertainty is something we need to
live with, and we think that we want to see a
treatment of uncertainty commensurate with its
importance to risk, and also a recognition that
information in the performance confirmation period
will continue to be collected and also will shed some
light on the uncertainty.
CHAIRMAN GARRICK: The issue is not to
eliminate the uncertainty, the issue is to quantify
them.
MR. McCARTIN: Yes, exactly. And what you
will see today is, really what I am trying to show is
that what we are trying to do is get a better
understanding of the uncertainties and possibly use
that to refine the agreements and help in our getting
ready to review a DOE license application.
(Slide)
With that, let me go to the first area,
which is the probability of an igneous disruption.
Basically, in terms of the issue, estimating the
probability of disruption is related to how many
volcanoes are currently there from the past, and
looking at subsurface geologic features. Those are
two parts of, we believe, in estimating the
probability.
In terms of the NRC approach, we're using
the currently identified features from the past, and
that is approximately 13 events over the past 11
million years, and using subsurface geologic features
to better constrain what might be the probability of
disruption at the Yucca Mountain Repository.
Nevada is using a smaller subset of that.
They are using primarily volcanoes that have occurred
over the last million years comprising two to three
events. They use that in the defined particular
zones, some narrow zones that tend to focus on the
Repository.
MR. LARSON: Is that Cole (phonetic) and
Smith, or is it Smith and Kenno (phonetic), or who is
it?
MR. McCARTIN: With that, I'll refer to
Brit.
MR. HILL: This is Brit Hill, from the
CNWRA. That's the series of models from Smith, et.
al., 1990, also a series from the Cole and Owen Smith.
MR. LARSON: Does this take into account
the recent work of Smith and his student Kenno?
MR. HILL: No, it does not.
MR. McCARTIN: Notably, the zones are not
constrained by subsurface geologic features. DOE's
approach, they are constrained -- their probability
probably primarily on volcanoes over the last five
million years, that's approximately seven events.
They just define some broad zones, some of which miss
the Repository.
They also are not using geologic features
to constrain the probability.
MR. LARSON: Is that really right? Is
that really right? It seems to me that DOE's approach
does use geology in the sense that they are limiting -
most of the source-zones are limited to crater flat,
and also they use topography, which is certainly, in
this area, part of the geological regime.
MR. HILL: Again, this is Brit Hill, from
the Center. When we say the zones are not bounded by
geologic features that localize volcanoes, we've
addressed the technical basis on why alluvium or small
topographic changes do not localize volcanoes, and are
restricted solely to the crater causation. And zones
that are defined based on features that do not
demonstrably localize volcanoes are the ones that are
in many of the source-zones that bypass the
Repository.
I would also point out, in the PVHA
report, when the experts talk about how they define
the zones, they clearly do not make a linkage to
exclusive geologic features.
(Slide)
Now, with that, what does it mean
quantitatively in terms of where the various groups
are at, and the horizontal lines on the left are
depicting the range of probabilities that were seen
estimated by the three groups. You can see, State of
Nevada-sponsored work is somewhere in the 10-6 to 10-8,
NRC is around 10-7 to 10-8, and the Department of
energy is somewhere -- a little bit of both, 10-8 down
to 10-10. The Department's mean value is approximately
1.6x10-8, which has it in that area of probability
space where it needs to be considered in terms of the
Yucca Mountain Repository, our probability cutoff
being approximately 10-8.
CHAIRMAN GARRICK: Now, are they all using
the same time interval? I thought I heard you say
that Nevada was using a different time interval.
MR. McCARTIN: Well, in terms of what are
the relevant events that they're using to determine
the probability, you're right. The easiest way to put
it, there's not an off-the-shelf thought to here's how
you determine --
CHAIRMAN GARRICK: So, these are not all
against the same database.
MR. McCARTIN: Correct. Well, people are
using different parts that they believe are more
relevant, and it is a matter of opinion. The
Department -- and I'll let Brit add anything -- but
the Department feels the most relevant events are
those in the last 15 million years, and they have a
basis for not including --
CHAIRMAN GARRICK: But this is such a
straightforward kind of question that we're asking
here, namely, the likelihood of the event. And it
would seem that you would not want to sweep any
knowledge under the rug that would shed any light on
the frequency of that event, of the recurrence of it.
MR. HILL: Brit Hill, from the Center.
First, a minor correction. The Department is heavily
focused on 5 million year and younger, not 15 million
year and younger.
MR. McCARTIN: Oh, I thought you said 15.
MR. HILL: I wouldn't characterize it as
"sweeping it under the rug" or anything like that,
except a difference of opinion by what is the most
relevant in determining recurrence in the next million
to 10,000 years. Whether you are using a tighter
range of information because of your understanding of
process, how these were formed by the same process or
not, or believe that only the last million years is
the most important for the next 10,000 years --
CHAIRMAN GARRICK: Of course, the
discriminator here would preclude the need for asking
the question would be if we saw the probability of
distribution, we'd be able to see what the shorter
time interval meant in terms of the uncertainties
versus the longer time.
MR. HILL: Right, in trying to gauge a
model that's based on recurrence rates of 1 to 2
events per million years, and you think about space
and time, and then look at -- you're trying to focus
on a 10,000-year interval which is so much shorter
than anybody's recurrence rate. I can give you a
sense of why, for example, our position is that really
an order of magnitude construction of probability is
about all that's warranted because both the time
interval, sparsity of data, and the long recurrence
doesn't warrant a large significant figure.
CHAIRMAN GARRICK: Well, the only thing is
that if it comes down to -- if it's a 10-7 value
that's important, and a 10-8, or whatever the line is,
is not important, then it becomes more important to
turn up the microscope a little bit on the different
assumptions that are made about the input information
for the distributions.
MR. HILL: Just one final point. It's not
really the input assumptions that are driving this,
it's the alternative conceptual models. And whether
we want to take a single tendency from a series of
alternative conceptual models or not is not at all
clear at this stage.
CHAIRMAN GARRICK: Well, I didn't mean to
spend so much time on it, but I just wanted --
MR. LARSON: There really has been very
little -- very few attempts at this kind of prediction
until this problem came up, and the predictions were
either on a very short-term basis -- an ad hoc basis
of days, hours, months -- and the longer-term like
we're talking about has just not been approached. So,
we don't have standardized procedures -- I think
that's what Tim said, that we don't have standardized
procedures. And so while we're doing this job, we're
trying to find those procedures.
CHAIRMAN GARRICK: The only thing I can
think that comes close to this is the frequency of
occurrence of very strong motion earthquakes.
MR. LARSON: Absolutely. There's a lot of
analogy there. In terms of short-term prediction
versus long-term prediction, these kinds of things,
this is the same problem, except that you have more of
an historical record with seismicity which gives you
a better chance of doing a decent job. And, also, we
understand that earthquakes are associated with
faults. Frankly, we have a really difficult time
pinning down what's really happening in the mantle
that's generating volcanic magmas.
Cognizant of your preamble, Tim, about not
getting into too much detail, but also recognizing
that you are talking about uncertainty, it strikes me,
when I see this diagram, I can recall during the PVHA
very lucid discussions of the NRC approach were made
by Brit and by Chuck Connor of the Center, and John
Trapp. I don't know whether I attended those or not,
it's been too long ago, but if I understand correctly
the results here and your previous slide, the ten
experts in volcanology that were selected by the DOE
to run -- to be involved in this expert opinion, chose
not to accept the approach that is being used by NRC
today.
Can you understand the uncertainties that
develop as a result of this? Can we understand why
they failed to accept this? Does this cast any light
upon the uncertainties that we see not within NRC's
work, but between NRC and DOE?
MR. McCARTIN: Well, in terms of between
NRC and DOE, I think it's -- in terms of -- I wasn't
involved in the PVHA either, so I don't want to talk
to that. I'll ask Brit to comment, if he'd like.
But, generally, this diagram here is showing the
variation of probability spatial based on the NRC
approach, and you can see the Repository loosely is in
this 10-7-10-8 range, and I think, obviously, these
aren't very -- you see variations in these isopleths,
and I think you're seeing the effect of structure,
which I think is useful in terms of new information
that could come in with the aeromagnetic data, but
you're seeing -- and I think Brit's right, in order of
magnitude, somewhere between 10-7 to 10-8. From my
standpoint, I look at the DOE current number, 1.6x10-
8, they are in that same range -- I don't know if one
could say we're in that big a disagreement at this
time. And as I said before and you indicated, there
is no textbook on how to do this, and there is
opinion, there is what-features-matter, et cetera, but
I don't know if Brit has anymore of what happened at
PVHA.
MR. LARSON: I just wanted to make the
point that -- and I do respect the NRC's approach and
people that are doing it very much, I respect them
very much, but I'm also struck by the fact that the
expert elicitation seems to disregard the NRC
approach, and I wonder where this is going to end up
when you come to some kind of judicatio on it.
MR. McCARTIN: One thing that at least I,
from my perspective and more PA rather than --
certainly, more PA than igneous -- but I think, as my
next slide will allude to, we're expecting to get new
aeromagnetic data that has been alluded, that there is
possibly 13 more identified events in the area.
When I look at what I've heard and my
limited knowledge --
CHAIRMAN GARRICK: Thirteen more, did you
say?
MR. McCARTIN: Yes, potentially --
potentially. It's still being evaluated, but it is
possible. Now, remember, you've got to look at the
time period, too.
MR. LARSON: Yes, location and the timing
is everything.
MR. McCARTIN: Is key, yes. But I think
when that information comes in, if I had to guess,
because of the fact that we use structure, subsurface
geology, to help constrain these numbers, I don't
believe the identification of those events will have
as big of an effect on the NRC estimate as it will on
potentially the State or DOE, and that's purely from
my very novice opinion, and does the NRC have a more
stable estimate because we have this other constraint
there that allows for -- this additional information,
depending on where it is, there are some bases. And
I don't know if Brit has --
MR. LARSON: But there was more than just
geological structure, I think it also had to do with
the length of the igneous dikes that were going to be
associated with it, and perhaps also the number, is
that correct? There's a difference between the NRC
approach and the PVHA approach.
MR. HILL: Again, this is Brit Hill, from
the Center. There are a number of questions that have
been raised. Going back to why the Center approach
wasn't used, I think it's just important to clarify
that the models that you see here that have geologic
structure incorporated into the clustering algorithms,
those models were not developed by the PVHA, they were
subsequent to it. So, this really represents our
preferred approach was not available at the time the
DOE conducted PVHA in 1995, so there's been a
difference in the available models, first of all.
Second, to clarify on the aeromagnetic
anomaly -- of course, that is fairly new information
-- the U.S. Geological Survey just released until it
could file a report, an interpretation that says that
these anomalies that are shown on the figure in the
labeled letters are -- can be interpreted as varying
basalt. It's not just the location of those features,
but the age of those features that's going to affect
all models, not just the State and DOE's, but it may
affect ours as well. All assumed occurrence rate is
fairly uniform and doesn't cluster in time. Until we
have a better handle on the age, it's very difficult
to say what the effect could or could not be. But
taking a recurrence rate of seven volcanoes over 5
million years to 20 volcanoes for 5 million years
could have a significant effect on any probability
model.
MR. LARSON: Can I ask another question?
CHAIRMAN GARRICK: Sure.
MR. LARSON: Going to this aeromag survey
-- and certainly it very much bears on the uncertainty
-- has the NRC considered the number of possible
basaltic features that might have been uncovered if
that survey had been flown with the specifications
appropriate for a survey designed for detecting these?
This aeromagnetic survey was conducted to map
hydrologic features. It was flown in a direction in
which you cannot adequately map dipolar features that
are small -- and I've written to you about that in a
report -- and, also, if this is so important to
determining the igneous concerns, why isn't this
magnetic survey extended into other regions, to the
north and to the east? There is a 300,000 year old
volcanic feature to the northwest, off of -- called --
MR. HILL: Thirsty Mesa.
MR. LARSON: -- Thirsty Mesa, thank you.
This survey doesn't go up into that area. If this is
so important, why aren't you pushing to get
appropriate data to do the job?
MR. McCARTIN: Once again, Brit will have
to answer that. However, what we're trying to do here
is give you a sense of the differences between
ourselves and DOE, and we certainly can go into that
detail, but Brit is going to have to answer that
question.
MR. HILL: Well, we had looked at
available information at the time, and thought that
the Cain and Bracken (phonetic) survey, the existing
aeromagnetic data, we had evaluated that and, as you
are aware, conducted a series of ground magnetic
investigations to look for potentially varying
characteristics, and we felt that with the available
information we didn't have a significant concern. Now
we're seeing new information from this aeromagnetic
survey. And I agree, it was not flown designed to
look for these features, but it was a target that
provided the Department a chance to mine the existing
data and come up with new characterization
information. Now we have the report, and it's raising
some very potentially significant concerns about how
many features, varied igneous features -- it doesn't
mean intrusive, necessarily, it could be buried
volcanoes -- really are within an area that could
affect our understanding of the probability.
MR. LARSON: Does that open up then other
areas for further investigation based upon what you're
seeing here?
MR. HILL: We're going to have to have
discussion with the Department of Energy fairly soon,
once we've had a chance to evaluate the open final
report that came out last week, but this is the
preliminary interpretation, which is also part of the
technical basis impact weather report where the
preliminary interpretations were that there were
additional features, but they would not affect
Department of Energy probability models.
MR. LARSON: Well, if you find all these
to the south and west of the Repository, I just wonder
what's up there to the north and west, because we do
have one that certainly falls within the time span of
anyone's evaluation of the volcanic processes in the
area.
MR. HILL: We're very concerned about
noisy terrain both adjacent to or east of the
Repository as containing present but undetected
igneous features.
CHAIRMAN GARRICK: Tim, this is something
-- do these numbers within these contours, say, the
10-7 to 10-8, do they reflect your uncertainty about
the likelihood, and is that all informational as
opposed to modeling?
MR. McCARTIN: Well, in this area,
certainly these are related to models that the Center
has used, and there's variation within those models.
In terms of whether modeling uncertainty or
uncertainty of another kind, I think Brit has to talk
to variation there.
MR. HILL: A very good answer is that the
uncertainty in the parameters, and also alternative
conceptual models. It does not capture the model
uncertainty. We have not evaluated quantitatively the
uncertainty in each probability model and factored
that in as an uncertainty.
CHAIRMAN GARRICK: Right, because this
suggests a pretty high level of certainty.
MR. HILL: We did not overinterpret each
line being a solid line that's fixed to the ground.
CHAIRMAN GARRICK: That's the only reason
I asked the question, don't want to do that.
MR. LEVENSON: Tim, I have a question.
Just because somebody wrote the words down forces me
to ask the question. In the Center review of DOE's
elicitation of expert opinion was criticized for what
ten people said greater balance is needed on the panel
to encompass a wider range of viewpoints, and also
potential conflicts of interest. Is the same
criticism potentially applied to inhouse version?
You'd probably have less than ten people.
MR. McCARTIN: Well, we certainly had less
than ten people. That's -- I'm not sure how to answer
that question.
MR. LEVENSON: I wouldn't have raised the
issue, but you people raised it.
MR. McCARTIN: Right. I mean, there's a
difference in looking at DOE, what they need to do to
provide this information. We certainly are working to
develop our own understanding here -- and I'll get to
you in a second, Brit, I know he wants to add
something here.
We have gone for peer review and outside
review of what we've done. We certainly have not done
a separate expert elicitation.
MR. HILL: We have not held our work out
as an expert elicitation. We put it as expert
judgment, and we put it in peer review literature, but
we have not tried to portray this as a consensus.
MR. LARSON: Well, I don't want to come to
anyone's defense here. God help me, I don't want to
do that. But I sat in on some of the reviews of --
for the peer reviews of the Center's work, and
although there were problems, there was general
consensus that this was really headed in the right
direction.
In addition, when many of their
publications came up for review, I happened to be
editor of the Journal and they fell in my lap, and I
don't think I've ever given anything as tight a review
as the papers that came in from the Center. Instead
of having two, three reviewers, they had five or six,
including members of the DOE staff. So, I think that
this has been looked at by the public and been pretty
well received.
CHAIRMAN GARRICK: Is that articulated in
your database and in your supporting information, what
Bill just described?
MR. HILL: I'm not sure what you mean by
supporting information.
CHAIRMAN GARRICK: Well, that's pretty
valuable evidence, it seems to me, what he just
described.
MR. HILL: It's not part of the record.
MR. McCARTIN: In general, though, our
approach has been to make sure DOE -- what information
do they have to bring forward to support so that we
can review their license application?
(Slide)
Continuing -- hopefully I'll be able to
pick up the pace.
CHAIRMAN GARRICK: That's not your
problem.
MR. McCARTIN: In terms of probability,
we've pretty much discussed this quite a bit. As I
said, the new aeromag data is coming in. From a risk
standpoint, what we see now is there's an
approximately an order of magnitude difference in the
probability between ourselves and DOE.
MR. LARSON: DOE keeps talking about their
probability assessment as being very robust, and I
think perhaps if we asked you, you'd say the same
thing.
What's going to happen with -- let's say
that just a few of these anomalies are volcanics that
fall within a range of distance and time that are
interesting. What's this going to do to 10-7, 10-8?
What are we worried about? Are we so -- is this whole
approach so fragile that finding another volcano or
two is going to send this catapulting to an even
higher probability?
MR. McCARTIN: Well, as indicated here, if
you ask for a best-guess at the moment, we're looking
at a 2 to 5 times increase in probability.
MR. LARSON: On the basis of 13 more
volcanoes in their position, in their place, as you
have them? So you analyzed them?
MR. McCARTIN: Well, what we're trying to
do is give you where we're at today and where we think
things may end up. And to say -- I mean, I don't
think we want to be held firmly to anything here, but
trying to give you an informed estimate that's where
we think it might end up.
Is that a big deal? I think what I'd like
to do is look at all the uncertainties we have and
look at it from that standpoint.
MR. LARSON: But, Tim, this is really the
critical uncertainty. The DOE oftentimes approaches
this, and many people approach this, from the
standpoint that you're multiplying zero times any
number is still going to be zero, and such a low
probability -- the probability is the backbreaker on
this.
MR. McCARTIN: Well, in one sense. I
don't believe that's so much the case currently, and
for this reason, that if it was below 10-8 and now
we're looking at a decision to whether it's considered
or not considered and it's screened up, that's not the
case. Right now, DOE is, like I said, approximately
1.6x10-8, so it's being considered.
And at least from my standpoint, not
looking at whether it's probability or some other
factor, if the estimate -- clearly, if the probability
increases by a factor of 5, it does increase the
overall risk by a factor of 5, but there are many
other aspects of the calculation that have similar
effects, either raising it or lowering it. And so
this is one of those factors, but if it were not in a
position of whether it's considered or not considered
-- and that's where I think the critical aspect was.
Whether it's 10-8 or 10-7, I don't think is as critical
as it's going to be considered or not going be
considered, at least from my perspective.
CHAIRMAN GARRICK: Tim, we're sorry to
press you the way we are.
MR. McCARTIN: No, no. It makes it more
interesting for me.
CHAIRMAN GARRICK: Let me ask this
question. Supposing that the TSPA/LA comes out with
a much more robust analysis that shortens the time to
the peak dose dramatically, as well as the magnitude,
and we already see scenarios where the peak dose time
varies from a few thousand years to a million years,
and it is swung dramatically by certain parameters
and certain performance characteristics, such as
actinide solubility and -- so, if it turns out that in
the license application or in the analysis that's
going on now that's a follow-on from the SSPA, that
this dose drops down and the time at the peak dose
shortens dramatically, does that impact the
significance of the 10-7?
MR. McCARTIN: In terms of -- you're
saying if the base case dose shortens and --
CHAIRMAN GARRICK: Yeah. It's one thing
to talk about recurrence of the order of 10-6, 10-7,
when you have a situation where the worst part of the
problem is of that same order in terms of time, but
it's another thing when that time dramatically
shortens as well as the magnitude of the dose. And
I'm just asking if the performance calculations change
dramatically, does that change the significance of the
recurrence interval for igneous activity?
MR. McCARTIN: Well, I guess I still have
to ask, with respect to the base case, you're saying
that the base case scenario occurs much earlier --
CHAIRMAN GARRICK: Yeah.
MR. McCARTIN: -- so that there's not --
CHAIRMAN GARRICK: I'm saying if they come
forward with --
MR. McCARTIN: -- would they add together
rather than be separated in time?
CHAIRMAN GARRICK: -- yeah, one in ten
million year event, how significant is that when now
you're talking about something of the order of 103rd
years and a much lower peak dose? It seems to me that
you've got to reassess the significance of --
MR. McCARTIN: I think that the
information that DOE will give us, we'll be able to
see what the base case results -- what the disruptive
ones are. The Commission will have to weigh that in
terms of what does that mean, if these two are
additive and now they are slightly above the standard.
CHAIRMAN GARRICK: Now, picking up on
something that Bill said earlier alluding to the fact
that the DOE has come forward -- feels that their
recurrence interval calculation, their probability
calculation is very robust, and yet they have kind of
caved in to the NRC number. Does that make sense?
Have they been able to -- have you been able to
demonstrate to them that your number is comparably
robust and that's why they chose to do this?
MR. McCARTIN: Well, in this particular
area, there is an agreement that the DOE can come
forward with a license application with their
probability number.
CHAIRMAN GARRICK: Oh, I see.
MR. McCARTIN: They have agreed to also,
as the sensitivity analysis, to include analysis with
the 10-7 value, to provide that additional
information, but it's sort of a two-prong kind of
thing.
CHAIRMAN GARRICK: Okay.
MR. HILL: Just perhaps one very brief
point, the reason that we have that agreement is
because at the staff level we had fundamental
disagreement with the robustness of DOE analysis.
CHAIRMAN GARRICK: That's what I was
curious about.
MR. HILL: In August of 2000, we had a
technical exchange that seemed like we were really in
a fairly impractical position on several of the
issues, and determined that if we got the analysis as
part of the TSPA/SR and TSPA/LA we agreed would have,
in our view, an adequate basis at 10-7 at that time.
That would provide us with enough information to reach
a licensing decision, and we would not need to move
the issue forward, but it's more that this was to
address specific technical concerns at that time.
CHAIRMAN GARRICK: See, I have a --
MR. HILL: It was not a matter of we were
just insisting on this number for some unspecified
reason.
CHAIRMAN GARRICK: Yeah. See, I have an
ulterior motive here. I'm trying to extract from this
whole process how much of it is compliance-driven and
how much of it is risk-driven.
MR. McCARTIN: In terms of the path
forward, there's really two parts of it. The second
-- that last part we've talked about enough, the 10-7
value that DOE will use, but the first part -- I think
both ourselves and DOE will be analyzing the
aeromagnetic data and updating estimates, and I think
including uncertainty is a key part of that, that I
think one of the things we'll be looking at, does it
make sense to have a single value, or should we be
sampling from a distribution --
MR. LARSON: Does that mean that DOE will
go beyond the report that was published this year, of
the USGS? Is Los Alamos continuing their work on
this?
MR. McCARTIN: Well, right now DOE, I
think, is updating a lot of their plans and what they
are going to accomplish, and when. And I don't know
if I want to try to speculate what DOE is going to do.
I assume they will analyze this information, but in
what time frame and how, I don't know. We would
expect to analyze -- to speak more for ourselves, and
I put 2002 there -- we're going to analyze this
information this year. I would expect DOE to do the
same but, like I said, they are updating their plans.
(Slide)
I finally got off probability.
CHAIRMAN GARRICK: That was the easy one.
MR. McCARTIN: Uh-oh, I'm in trouble. In
terms of getting in more now to the consequences in
terms of the Repository magma interactions, the issue
is one of when magma rises up and there is a drift,
there's obviously a path there that it can take down
the drift, and the question is one of what's going to
happen in terms of an alternative pathway. Rather
than continuing straight up to the surface, the magma
goes down the drift and breaks out at some more
distant part, thereby intercepting more waste
packages. And so it has to get to how many waste
packages are intercepted in the event.
Alternatively, the NRC approach, we still
have a single vertical conduit, 1 to 10 waste packages
are intercepted, it depends on the diameter of the
conduit, on average 5 waste packages. However, there
are these numerical experiments and calculations being
done to look at what happens when the magma hits the
drift, and there are evaluations for that at the
current time we are looking at it is on average. It
could be as high as 100 waste packages, depending on
some assumptions in terms of how far down the draft
things go. But if we represented today this
alternative, you'd be looking at approximately 100
waste packages versus the 5.
DOE has a very similar approach right now
of looking at a single conduit up through the
Repository. They sampled somewhere between 3 to 30
waste packages, a median value of around 10, as part
of this vertical conduit. They have agreed to look at
this alternative approach. We haven't seen any of
their analysis related to this, but that's part of the
--
MR. LARSON: Is the alternative approach
only for dikes, or is it for a pipe?
MR. McCARTIN: Well, if the dike comes up,
intersects the drift, the Repository drifts, and then
the question, where does the conduit form? And it's
the dike actually that's intercepting the drift, and
could intercept multiple drifts, which is how we get
the possibility for 100 waste packages, and it is on
average 2 drifts are intercepted and where conduits
form.
MR. LARSON: Maybe the committee knows
this already but, if not, just what is NRC doing in
this area at this time, have you people learned that?
Do you know what --
MR. McCARTIN: Well, I thought in August
Brit Hill gave a presentation to the committee in
terms of the -- both there are numerical calculations
going on with laboratory experiments, analog
experiments, to try to corroborate the numerical
calculations. I don't know if Brit wants to add to
that at all, but --
MR. HILL: We also went over this last
month.
CHAIRMAN GARRICK: The one thing that has
always been difficult for me to understand, and not
being an earth scientist, it's not difficult to
understand why I don't understand, but it's one thing
to get these recurrence intervals on the basis of the
kind of data we've been talking about, but what do you
use to establish your insights on pathways and depths
and magma flow and what have you? Where do you get
that information? How do you get that information?
MR. HILL: There's a lot of sources of
information. At first, we are strongly based on an
analog approach where we looked at similar volcanoes
that have been recently active. Second, on the flow
pathways, we're taking a stronger numerical and analog
experimentation approach because we can't go down 300
meters and look at these sorts of conditions. And
it's unprecedented that ascending magma has
intersected a horizontal void of any extent at these
kind of depths. We hear a lot of tubes, but they are
very, very shallow, and so you don't have the same
sort of depressurization and flow phenomenon that you
would expect down at 300 meters. That's in part why
we've been doing a lot of the experimentation and are
talking about this year continuing with fluid dynamic
flow experiments to better look at this kind of a
process at the appropriate scale for fluid containing
a lot of gas. But there really is a lot of
information out there on how normal volcanoes of
similar composition and volume and character, how the
volcanoes erupt in nature. We can glean an awful lot
of the physics and fluid dynamics just by very
straightforward observation and simplified modeling.
And, remember, we're talking about a pressurized fluid
intersecting an atmospheric void.
CHAIRMAN GARRICK: Right. And I guess the
issue of backfill comes into this in a significant
way.
MR. HILL: Certainly.
MR. McCARTIN: From a risk standpoint,
obviously we are looking at an approximate order of
magnitude, when you look at 1 to 10 waste packages, of
up to 100 on average.
In terms of what's going to go on to move
the path forward, DOE -- we're expecting them to
analyze this scenario and come to some conclusions.
From NRC's standpoint, as Brit indicated,
we are continuing with the numerical analyses in some
of the experiments. In September of this year, we
expect to have some additional information on
verification of the model and the experiments, and as
you indicated, backfill, and that's exactly what this
last tic is, looking at the consideration of
Repository design in terms of how this impacts the
number of packages that could be affected. Backfill
certainly has a big impact. Right now, the design is
not to backfill the drifts, but the access tunnels, et
cetera, would be backfilled, and that has some
implications.
MR. LARSON: Are you having any peer
review of the work that you are currently doing,
verifying the numerical models and the analogs, the
laboratory analogs?
MR. McCARTIN: A couple of them. Well, I
guess it was about two years ago we had -- well, some
might call it a peer review, we certainly brought in
a number of different experts in different areas to
review what we were doing with the TPA code, and got
review comments, et cetera -- maybe an expert review
rather than a peer review. And certainly the Center,
as a group, tries to publish in a number of journals.
NRC staff did the same thing, and so we continue to
publish. I don't know if there are any explicit plans
for any specific review of a particular topic.
MR. HILL: The publication is to elaborate
very briefly. We've been having trouble getting the
two reports that we talked about earlier last year,
getting those reports accepted for review because the
topic is deemed too esoteric to appear in Geological
Journal. So, we have yet to receive --
CHAIRMAN GARRICK: I knew these scientists
were stuffy.
MR. HILL: But we're continuing to
resubmit to different journals that have a little more
dynamic approach. Of course, this has been presented
at international meetings like the American Geological
-- or American Geophysical meeting.
MR. LEVENSON: Tim, I have a question in
the context of trying to move toward risk-informed.
There's a statement in the report here that "Staff
concludes that the character of past YMR igneous
activity represents the most conservative bounds on
future YMR activity". That most conservative, 1 order
of magnitude, 3 orders of magnitude?
MR. McCARTIN: I guess I'm not -- what
report is that?
MR. LEVENSON: That's the Center's report
on --
CHAIRMAN GARRICK: It was in our briefing
book.
MR. LEVENSON: -- our briefing book. I do
read the junk you send us. It's Technical Basis for
Resolution of Igneous Activity.
MR. McCARTIN: I'd be happy to take that
page down and -- I can't --
MR. LEVENSON: I'm really asking a generic
question. I get nervous whenever people say, "Well,
that's very conservative, so it's okay". I mean, is
it a factor of 2, is it a factor of 100?
MR. McCARTIN: Well, partly, what I'm
trying to do in going through this is give sort of
where things could end up, and I will talk to that,
but the approach we're trying to do and part of what
this is as a result of, we're looking at the
information that we have, what we're using to estimate
the consequences, and how we might improve things to
get a more realistic approach.
We certainly are not trying in any area of
the TPA code to take the most conservative approach.
I'm not -- those words trouble me, I guess, because
it's not clear on the context --
MR. LEVENSON: It isn't clear from here
whether it's the NRC staff that did that analysis or
what, you can't tell who did that. This is the staff
commenting on it.
MR. SINGH: I'm not sure if that's been
issued to -- sounds like it's still being reviewed.
MR. LEVENSON: I'm sorry, go ahead.
(Slide)
MR. McCARTIN: In terms of the magma-waste
package interactions, clearly, when a hot magma
interacts with a waste package, we're talking about a
-- certainly, in the conduit, a fairly violent
interaction. Physically, chemically, thermally, these
conditions are quite extreme.
So, the question is how does the package
act in this environment? In terms of the NRC, for the
extrusive amount, the package is in the conduit.
We're assuming the package offers no protection from
the spent fuel from the magma. And so it's as if
there is no package.
MR. HILL: Just to answer the previous
question about the character of Yucca Mountain
volcanoes conservatively, that's specifically to
address that we do not believe that magma water
interactions would create a different class of
volcano, and that there is no other class of volcano
of this composition that would give a more dispersive
eruption, which is the process that was discussed in
that section. So, the past character would bound the
dispersivity and fragmentation capability, and that
any other -- trying to say that they're less
dispersive would be a less conservative approach.
MR. McCARTIN: In terms of the intrusive
amount, we have on average in our code approximately
40 waste packages failing, and just assume, once
again, if magma contacts those waste packages, it will
fail the container. Part of it is due to the
temperatures. The magma is approximately 1100
degrees, et cetera, and the drifts are unbackfilled.
This really does not account, at this time, for this
alternative flow path. It could be more packages if
we looked at that alternative flow path. We have not
worried about varying that. We believe that's a very
small contribution to the overall risk, and you'll see
that actually is a slight difference between ourselves
and DOE.
In terms of the DOE approach --
MR. LARSON: Excuse me, Tim. Is that then
all taken -- enters into the biosphere through
groundwater?
MR. McCARTIN: It has that possibility,
yes.
MR. LARSON: But only --
MR. McCARTIN: Only -- yes, it is
intrusive, yes, absolutely. For the DOE approach,
they have a similar approach for the extrusive. Any
waste packages in the conduit don't have any effect of
limiting the entrainment of spent fuel in the magma.
In terms of the intrusive, they have a
slightly different approach. They have two zones.
Zone 1 where the package offers no protection, also
similar to ours, and in that zone right now they have
approximately 200 waste packages. And then there is a
Zone 2 where they have what's called some "end-cap"
failures, some moderate failures of the waste package.
And they have, on average, 2,000 waste packages there,
a significant amount more. And that is a big
difference between ourselves and DOE. They have far
more -- they get a larger intrusive release than we
do. However, even if we increased our number of waste
packages, we would not get, I don't believe, get to
the numbers that DOE has, and that's something that
you'll see in terms of the path.
CHAIRMAN GARRICK: Well, what they gain on
their probability, they lose on their consequences.
MR. McCARTIN: That's one way to put it,
for intrusive. That's the intrusive -- yes.
(Slide)
And that's part of the risk insight here.
We understand that the intrusive is a very low
fraction of the extrusive, and so we have not
concentrated much effort on that particular part.
Conversely, DOE has the intrusive a much
larger fraction. Until recently, the intrusive was
the larger dose contributor. In the SR, intrusive was
a larger dose than the extrusive. That has
flipflopped, but it is still a very high percentage of
the extrusive.
CHAIRMAN GARRICK: Now, have you seen the
analyses that they performed where they assume these
2,000 waste packages?
MR. McCARTIN: Yes, that's the SSPA.
CHAIRMAN GARRICK: So you're able to pin
it down as to the fundamental difference between the
NRC analysis and the DOE analysis?
MR. McCARTIN: Qualitatively, I believe
so. I'm not convinced. I mean, that's one of the
things, path forward, that this last -- this year, I
promise you, we will understand those differences. I
think they're -- at a very broad level, I believe
there is approaches in terms of the release, in terms
of their diffusional release, that when they go to
this igneous scenario where they fail the waste
package, I believe they give some very, very large
diffusional releases, and that's part of it.
There might be some ways that they are
using the probability to weight things that we don't
quite understand how they are doing it. They have a
slightly different approach to weight the consequences
with a probability than we do, and there are some
other aspects to the calculation. We are digging into
it. We don't have the answer. This year, we will.
I think part of it is we need to get a little smarter,
look a little deeper into the DOE calculations, and
ultimately we will end up with probably an Appendix 7
or some other type of meeting with DOE to go over,
okay, here's how we understand your representation.
As you know, I've mentioned this before,
diffusional -- one simple thing to point to is
diffusional. Ourselves, and DOE, for the base case,
have approximately the same releases for completely
different reasons. We take no credit for cladding,
but we have no diffusional releases. We did an
estimate and saw that diffusional releases were such
a small fraction of the invective releases that we
don't have diffusional release.
DOE, on the other hand, takes a lot of
credit for cladding, and has a diffusional release
that dominates over their invective release,
approximately an order of magnitude, and I think we --
CHAIRMAN GARRICK: Particularly for long-
term effects.
MR. McCARTIN: Yes. And in the situation
for volcanism in the intrusive, I believe for those
200 waste packages they take no credit for cladding
and no credit for the waste package, and so they have
some very, very large releases. I think that's at the
heart of it, I'm not certain, but there's an aspect of
taking in and understanding this better that we
certainly are in the process of doing. I'm very
confident this year we will have an answer to that.
To me, from a performance assessment standpoint,
that's the most fun this job brings, is trying to
uncover a mystery. We have a mystery here we don't
quite understand, and technically it will take a
little bit of work, but I think that's the fun part.
CHAIRMAN GARRICK: Andy.
MR. CAMPBELL: Tim, are there any examples
of magma interaction with features at all? The reason
I ask is, if I understand correctly, basically, once
a waste package "fails", for all intents and purposes,
it is not there.
MR. McCARTIN: Right.
MR. CAMPBELL: And you haven't talked
about waste form interactions, but essentially you
assume a significant fraction of the waste form
becomes a very small particle. Is there anything that
can be gathered in terms of interactions with engineer
features or wall rock interactions that can constrain
these kinds of models?
MR. McCARTIN: Certainly, there's wall
rock interaction. On here, we are going to try to
look at the literature for analogous kinds of
situations, and it's primarily the Center is going to
be helping us, try to see are there things that we can
draw some parallels. It's not easy, and I don't know
if Brit has any other ideas, but we're going to try to
look at that in a data-poor situation and try to find
some additional information -- I know Dick Codell is
trying to look at are there some reactor accidents,
like TMI, where maybe there's some information more in
the fuel area, can you glean some information in an
area where you have very little, but we are going to
try to pull in information where possible.
MR. LEVENSON: There is data from
Chernobyl, both the U.S. models and the Russian models
predicted that the molten core would certainly go
through the floor down, and it didn't penetrate the
floor anywhere. The molten fuel ran for long
distances on the top of floors, and then poured down
through existing holes, and all of the models were
incorrect in their projections, both U.S. models and
Russian models.
MR. HILL: To directly answer your
question on engineering analogs, the answer is no.
There is no experience of having a basaltic volcano or
any volcano erupt directly through an engineered
facility. And, remember, we're talking about putting
a waste package into an erupting volcano, not on top
of a lava flow. So, there is a long-range of
physical, thermodynamic process that's involved in
this, and exactly modeling what's going to happen as
we go from highly reducing environment at temperatures
of about 1100o C. under very high dynamic flow, the
best I can tell you is that this starts off as a 1
meter wide dike and reams out the wall rock on the
order of 10s of meters in diameter, through
overpressure-underpressure relationships, so there is
sufficient force, sufficient work to essentially at
times melt or disaggregate solid rock.
And just a clarification on some of the
Chernobyl information, I believe the temperature for
the fuel to become molten is about 2000o Centigrade,
there's some very high temperature beyond what we have
been seeing in igneous events, or well below the
solubleness for incorporation in basalt, but we're
dealing with, again, a process where at the point of
initial incorporation where a highly reducing
environment of producing these are very low, on order
of 10-9 atmospheres, but as we go up into the erupting
cloud, we're going -- mixing with the atmosphere very
rapidly, at temperatures on the order of several
hundred degrees Centigrade. And one analogy we have
drawn from the Chernobyl accident is a very rapid
oxidation and formation of secondary oxide phase and
embrittlement of fuel particles during these complex,
rapid geochemical mixing events. We are continuing to
work with people like Dick Codell and other folks that
have a much better background in waste forms to try to
glean some things from these reactor accidents to look
at more mechanically how would the fuel aid during an
event, but we're dealing with very rapid changes of
very high dynamic load on all of the waste packages
waste forms.
CHAIRMAN GARRICK: Just in that
connection, what do you assume about the deposition of
the radionuclide inventory in the magma, is it just
instantly available?
MR. McCARTIN: There's an incorporation
ratio that picks it up.
MR. HILL: And we're looking at a process
of conduit-widening right now where the waste package
is heated up for ambient conditions up to about
magmatic temperatures, and the internal contents are
heated as well, so we're going to have cladding
degradation, oxidation --
CHAIRMAN GARRICK: So there is a process.
MR. HILL: It's process-driven, but it's
not mechanistic because there are so many
uncertainties about how would this feature behave. We
faced early on the dilemma of, well, guess, it's not
instinct, would it be 100 percent efficient? Probably
not, except if I look at holes in the ground from
volcanoes, they're 100 percent efficient in
lightening. But how would I get a technical basis to
say it's 80 percent or 70 percent efficient? How
would I defend that?
CHAIRMAN GARRICK: So it's just one of the
many abstractions that are involved.
MR. HILL: It's a complex abstraction,
that is a simplification, but it is based on the
unusual mass and mechanical loads that are inherent in
this system, analogous for a reactor loci than any
sort of storage type accident.
CHAIRMAN GARRICK: Thank you.
(Slide)
MR. McCARTIN: Going to the fuel, along
the similar lines, but certainly once again the
chemical/thermal aspects of the magma are going to be
a harsh environment for the fuel. Right now, NRC has
a fairly simplified approach -- it's an incorporation
ration in that particles -- in simple form, particles
that are larger than the spent fuel particles can
incorporate that smaller particle.
DOE has adopted a similar approach.
MR. LARSON: Does that take into account
the density of the -- the high density of the fuel?
MR. McCARTIN: Yeah.
(Slide)
In terms of the risk insight, there's a
lot of uncertainty here, as Brit was describing,
exactly how is this fairly violent eruption, as its
pushing through the conduit, interacts with the fuel.
Certainly, you can do some further refinement, but our
gut feeling is that the refinement would not result in
a significant change. Right now, 100 percent of the
fuel is incorporated. Is it 80 percent? Is it 50
percent? You might have to do a lot of work to get a
little more resolution on that, and we're not
convinced you could get it down very low, but that's
a gut feeling. Certainly, our path forward this year,
we're looking at refining our source-term model in
that sense. Once again, as was mentioned, we're going
to try to evaluate the relevant reactor accidents
possibly. Maybe there's some information there that
will help us.
And the bottom line is, DOE does need to
develop a technical basis for this. This is one area
where I know they have adopted our model. Merely
adopting the NRC model is not a technical basis. Just
because it's our model doesn't mean it's necessarily
supportable. The DOE needs to do some work to at
least convince themselves that it is a reasonable
model, not just because it's the NRC's.
MR. LARSON: Are you evaluating more
complex repositories that simply are a tube? Are you
assuming any distortion of the tube? Any rock falls,
and backfill?
MR. McCARTIN: In terms of the alternative
pathways?
MR. LARSON: Right.
MR. McCARTIN: Yes. That was part of the
repository design, you look at backfill. But we also
are looking at, over time, you'll have rockfall and
accumulation of material in the drifts.
MR. LARSON: If I understand the process
here, what you would end up with is the canisters that
are being opened up being pushed to some portion of
the drift. Are you evaluating the effect of that, of
a concentration of the fuel on the thermal aspects of
the surrounding rock? You know, if we're going after
a low temperature repository, we'd push all of the
magma -- or all of the fuel into one end, we've
changed that process.
MR. McCARTIN: Yes. The initial
calculations were assuming empty drifts, and
subsequent ones are looking to refine things -- and
Brit can talk to the later work.
MR. HILL: Right. The initial scoping
report back in 1999 talked about the possibility of
the initial shock being sufficient to move waste
packages. In the Woods, et. al. paper, we've refined
that approach with a little more added, and do not
believe that you have either enough friction or
velocity to move a waste package.
MR. LARSON: So, a lot of simplifying
assumptions.
MR. HILL: We're not seeing things like
moving down the drift, the waste package remains
intact even under the range of flow conditions in the
Woods, et. al. report. So that was an initial model
that turned out to be --
MR. LARSON: So there will be no
concentration of the fuel in any portion of the
drifts?
MR. HILL: Not during the initial stages
of flow because you cannot move an intact cold waste
package under the conditions that we currently realize
as reasonably bounding the expected upper range of
flow during that initial impact.
CHAIRMAN GARRICK: You can't have it both
ways.
MR. HILL: Another thing is, we haven't
looked segregated flow during sustained eruption.
After the waste package is disaggregated, it's
possible you could get a concentration zone between a
low velocity and high velocity horizontal flow, but I
don't think that's likely given the turbulence
inherent in the system. However, we haven't analyzed
it, but plan to analyze it in the coming year.
MR. LARSON: Thank you.
(Slide)
MR. McCARTIN: Well, we finally got out of
the repository now. We have an ash deposit on the
ground, and the question is one of with this ash
deposit at the RMEI location, how is it going to
evolve with time? There are really processes that
could remove material. There's processes that can
bring it in from, say, flooding from the repository,
100-year flood comes in and washes some of the ash
from near the mountain down to the RMEI location.
Right now, NRC has a simplified approach.
I have "conservative" with a question mark. I think
the gut reaction is that this is a conservative
approach. However, having criticized DOE for doing
the same thing, what is your basis for saying it's
conservative? And I think we have to do more work to
understand that. It may very well be.
Right now, we're having the ash plume blow
in the direction of the critical --
MR. LEVENSON: Excuse me a minute, Tim.
Before you get to the ash moving, what was the
approach in deciding what the composition of the ash
was, the ratio of plution or fission products or what
have you, to ash. Is this concentration of mass
involved here?
MR. McCARTIN: Well, there's the volume of
the mass of the --
MR. LEVENSON: You've got a limited number
of containers that have failed, and a huge amount of
ash.
MR. McCARTIN: We're assuming a uniform
mixing. At the RMEI location, the material that's put
down is uniform within the ash.
MR. LEVENSON: And the difference in
density of a factor of 10 doesn't give you any
segregation at all.
MR. McCARTIN: Right now, we're not
accounting for those kinds of things.
MR. HILL: Please recall that it's being
incorporated into molten material, not adhering on a
solid.
MR. CODELL: This is Dick Codell, NRC. We
are working on an alternative model of the fuel ash
incorporation where we do take density into account,
and it may supersede the incorporation ratio model.
In this case, the fact that the fuel is much denser
than the ash could lead to more small dense particles,
which may behave differently once they are thrust out
into the atmosphere, although we don't really have an
adequate model for dealing with that atmospheric
transport.
MR. HILL: We did do a very quick scope
that even if we fail 10 waste packages, there still --
with a small volume tetra eruption, you're dealing
with .01 weight percent of high level waste in the
total amount of the eruption. So this is really a
trace component in the total mass core volume of the
eruption.
MR. LARSON: But it would seem reasonable
that you could get some stratification in terms of the
distribution of the fuel contained tetra. It just
seems logical that there would be a stratification
with the density --
MR. LEVENSON: Especially at the low flow
rates. At the relatively low flow rates, there's not
turbulent mixing as this goes down the drift.
MR. CODELL: Well, the true is, what's
likely to happen, though, is that the biggest ash
particles will take the biggest fuel particles with
them, simply because there's more of the bigger ash
particles. There's more mass in the bigger ash
particles, so it's not going to change density of the
mixture that much. My best guess is it's a relatively
small effect.
MR. LARSON: But calculations are being
made on this?
MR. CODELL: Yes.
MR. LEVENSON: Having spent a number of
years trying to mix things to get uniformity, and
knowing how difficult it is, I have trouble assuming
this kind of a thing. I mean, the flow down the tubes
is not fast enough to push the containers, and yet
you're assuming complete mixing of everything which is
coming long after the containers have failed. It
doesn't sound like a good assumption to me.
MR. HILL: Just to follow up with that,
we're considering that we're dealing with a very
complex gas-fluid mixture that's taking a very
irregular geometry. Even though we talk about
horizontal flow, it's coming up from depth vertically
as about 50 percent fragmented -- 50 percent gas
volume, 50 percent magma -- both horizontally and then
breaks and goes vertically again along something that
starts off as a 5 meter in diameter tube, but we have
every reason to believe that as the wall rock is
stressed, the over- and underpressure will be plucked
and scoured the way any other volcanic conduit is.
So, while we may develop a quasi-angular
flow regime, there's going to be backpressure within
the system and there's going to be angular collapse if
it has any reasonable analogy to a volcanic conduit.
Both will produce an incredible amount of churning and
turbulence within the deposit. So, it may not be
uniformly turbulent with continuous incorporation, but
for the duration of the eruption, something on the
order of three weeks, we'd be having repeated overturn
and convection within the system. It would be hard to
say that you would keep segregation throughout the
duration of an eruption.
CHAIRMAN GARRICK: Does the scouring and
the other phenomena that's taking place have a
significant effect on the density?
MR. HILL: There may be very transient
effects due to gas resorption during slight
overpressuring, but that would quickly be -- it would
only be in the change in the bulk.
MR. LEVENSON: Because the wall is almost
the same density as the magma. It's the fuel that's
a factor of 10.
MR. McCARTIN: In addition to having the
wind blow south to the RMEI location, the question is
how long does that deposit persist at the RMEI
location? And right now we have a half-life of
approximately 1,000 years for the persistence of that
deposit. That's primarily based on the fact that in
looking at analogous deposits, that they seem to
persist in an area for around 10,000 years.
CHAIRMAN GARRICK: What do the wind rows
look like at Yucca Mountain?
MR. McCARTIN: It's quite varied, and it's
approximately, I'll say, on the order of 30 percent of
the time due south. Now, the other part of this is
that we don't account for any movement in to the
dislocation from other parts, to the flooding, et
cetera. There's no remobilization. We are merely
subtracting. It is a simple approach. Like I said,
we're actually going to try and do a fair amount of
work this year to get a better sense of the
reasonableness of this approach.
DOE has partly a similar approach. They
are having the wind blow due south also, however, in
terms of how long does the deposit persist, I estimate
a half-life of around 50 years, so it doesn't persist
nearly as long. And I'm trying to understand a little
better -- they do not use the half-life concept, they
have a different kind of approach, but between the
erosion rates they use how thick the ash deposits are,
I'm estimating that -- I think I'm close, but I may be
off on that -- but you can see significantly different
half-life than the NRC. At the NRC, we're looking for
a much longer lived.
CHAIRMAN GARRICK: As I recall, you used
the same erosion rate for longer periods of time as
you did for initial periods, did you not?
MR. McCARTIN: Yes. I mean, it's a
constant.
(Slide)
In terms of what this might mean, I think
in terms of the redistribution, I think increasing the
removal rate, if we went to a shorter half-life for
the deposit, I think there could be possibly an order
of magnitude difference. I'm not convinced of that,
but it could be between the DOE and the NRC approach.
There are a lot of other effects that need
to be considered in that. In terms of path forward,
I think next year we're going to try to evaluate, if
we take into account the local wind effects. As you
were saying -- Brit was indicating this could be last
three weeks, a month. How does this deposition look
like with a varied wind pattern? I think it would be
useful to look at that. Knowing, of course, what you
get in one area, you take away. Then if you spread it
out more, you have to be a little more careful about
the redistribution over time. And that's why we want
to understand, I think, better -- the risk
significance this year is, how much does this
redistribution really matter, and do some, obviously,
sets giving analyses with the TPA code and see exactly
what some of those assumptions mean. There's a lot of
-- one of the key things you want to be very careful
with in any analysis, and especially the TPA code
where there is a host of interconnecting things,
getting much better on one aspect like, say, the wind,
and neglecting these other things, the confounding
sensitivities that you might create by getting very
sophisticated in one area and not doing it in another
area, and that's a very complex problem that's part of
this analysis. Certainly, I can get better in one
area, but in terms of the overall analysis, what have
you now done.
CHAIRMAN GARRICK: Sometimes this is
referred to as the "lamppost" syndrome.
MR. McCARTIN: Absolutely, yes.
MR. LARSON: At what depth do you no
longer consider the radioactivity? You get an
accumulation. What is the maximum depth of
consideration?
MR. McCARTIN: Well, we only look at the
top 3 millimeters in terms of what would be possible
to be in the mass load, in the particulates. Now, we
keep track of the entire blanket in terms of -- or
deposit for how long it persists there, but in terms
of what's available for an inhalable dose, it is, I
believe, is the top 3 millimeters. It's some small
amount.
MR. LARSON: Taking into account the
porosity of the tetra amount?
MR. McCARTIN: Yes. Now, DOE has a
slightly different approach. They assume all the
radionuclides are in the top centimeter. We do not do
such a thing. That's where we start comparing that
very carefully. We're looking at the -- it's an
average -- it's uniformly mixed through the entire
deposit. DOE takes the radionuclides, as far as I
understand it, and puts them all in the top
centimeter. So, there are differences. There's a lot
of subtle differences. I tried to get on the ones
that -- for now, we'll be doing some analyses in the
future.
Here's another area where we think we can
improve the understanding of erosion processes and
things. Certainly, some surface water hydrologists at
the NRC -- Ted Johnson has been out to Yucca Mountain
to help us better understand how things might change
with major floods, et cetera. And so we're hoping to
do some more work along that line. And certainly with
respect to redistribution, there certainly are analogs
for movement of even ash deposits. I know Brit and
some of this Center colleagues have been to Sierra
Negro where there's an ash deposit, et cetera. So
there is some information that hopefully we can
continue to --
MR. LEVENSON: Tim, before you leave that,
remobilization, in this same infamous report, it says
"The high level waste contaminated tetra fall deposit
will be modified by wind and water for many years
after the eruption can be transported away into the
critical group by wind and water following most future
eruptions". Are you analyzing multiple eruptions?
MR. McCARTIN: No. Once again, I'd have
to read the whole page. I recognize the one sentence.
We certainly only analyze a single event. I don't
know -- Brit, do you know?
MR. HILL: I was the author of the report,
I think I can clarify that. You consider a variable
wind rows for any future eruption most of the time,
unless you had a wind directed blowing towards the
northwest sector at an extremely high wind speed,
something on order of 10s of meters per second, you
would have an appreciable amount of tetra falling on
east-facing slopes that drain into the 40-mile wash
drainage system. So, for most eruption scenarios, you
would have, even if the plume is directed away from,
at that time, the critical group location, you would
still have tetra that would fall on slopes that would
feed into 40-mile wash and the potential grade leading
down to the critical group location. It's very
difficult to have an eruption for a probability at the
proposed repository site, and not have material
eventually end up in 40-mile wash, even if the plume
is 180 degrees from it.
MR. LEVENSON: Well, that I understand,
but that's not what this says. This talks about
future eruptions many years later.
(Simultaneous discussion.)
MR. HILL: I know we haven't appreciated
--
MR. LEVENSON: Maybe it's just the words
that are here. I'm just reading what's here.
MR. LARSON: Is DOE conducting similar
studies?
MR. McCARTIN: Certainly, they have
evaluated the redistribution and the erosion at the
location of the RMEI. They right now are -- I don't
know in terms of analogs off the top of my head, I
know they are using USDA numbers for similar kind of
areas to get a general erosion rate.
MR. LARSON: There was some talk about
using Sunset Crater at one point.
(Simultaneous discussion.)
MR. McCARTIN: Could be, I'm not familiar.
But it is an aspect of the calculation that certainly
we think has some importance.
(Slide)
Finally, you end up getting a dose, and
the inhalation scenario generally is related to how
much dust or ash mass is in the air. Estimating the
mass loading as uncertainties, there are assumptions
about outdoor activities, et cetera.
For the NRC approach, once again, we have
-- and I tried to put these in similar terms between
ourselves and DOE for ease of comparison. We don't
use an average over 10 years, but DOE's numbers were
presented in that way, and so it's slightly easier.
As you can see, we have approximately on
average about 1.5 to 2mg per cubic meter. How did we
get that number? There are three components to it.
The first one is a high disturbance, and that's
looking at activity such as farming and plowing where
a lot of dust is raised, possibly traffic on roads, et
cetera, and there's a certain exposure time to that.
We have approximately 1 percent of the time in this
high disturbance type of activity.
Next, a lower value of mass loading for
general outdoor activity, being outside, walking,
other types of things. That exposure time is around
20 percent, and then sort of a background level that
is at around -- add the two, it will be around 79
percent to get to the 100 percent.
Generally, the dose is dominated by the
first two. Those two contributing about the same. You
can see right now you have about an order of magnitude
higher mass loading for this value, but the exposure
time is about an order of magnitude less. So, between
the two, the overall dose is really dominated by those
two.
In terms of, well, how long does it stay
dusty? What we do, we have a background mass loading
that is significantly smaller, and we have a half-life
of 10 years, and so this higher mass loading decays
into the background with a half-life of about 70
years. So, at approximately 70 years, it's going to be
pretty much at that background, but it gradually
decays. And that is an assumption. Some of that is
based on analog information at Sera Negro. That
deposit has been there for a while. How long will
this ash stay there in a fairly --
CHAIRMAN GARRICK: So what are some of
these conditional dose rates?
MR. McCARTIN: It depends on when -- well,
it's dependent on many things. Probably the two
biggest variables to give you would be -- one would be
the time that the dose occurs, primarily because of
decay of some of the key -- the short lived nuclides
that you don't see in the groundwater pathway, but if
you have an event at, say, year 100, it's certainly
more prevalent there rather than at 500 years.
And then how many waste packages do you
assume are entrained -- and it's always dangerous to
go off the top of your head, however, I will try to
give you my best estimate. I believe that for our
base case, it's on the order of 10-to-100 rems at 100
meters. So that would be the worst event, and that's
using these assumptions and everything else.
MR. HILL: This is in the IRSR Rev 2, and
Tim is correct, at about 100 years it's on order of an
average of 100 rems. At 1,000 years, it's on average
of order 10 rems, and by 10,000 years it's on average
of 1 rem as the conditional dose in that given year.
But it is in IRSR Rev 2.
CHAIRMAN GARRICK: Of course, if you were
talking about a 100-year compliance period, you'd be
talking about a substantially different probability.
So, the weighted risk is still the way you have to
look at it. But I was curious. Thank you. I wanted
to know what those numbers were.
MR. McCARTIN: There's been a lot of
evolution of this calculation, and even in the IRSR
Rev 2, we have changed -- some of the mass loadings
have reduced since then, and so there's a lot of
thinking going on, but certainly we're in the right
ballpark.
MR. HILL: I don't think we've had order
of magnitude changes, but again an order of --
CHAIRMAN GARRICK: Well, I think the way
you've stated it in terms of mg/m3 is the way to start
this process.
MR. LEVENSON: Do you have an estimate for
how much solid material is deposited in the lungs if
you're breathing 4.5 to 9 mg/m3? Do you die of the
dirt in your lungs?
MR. McCARTIN: No, no, it's not that high.
And be aware -- and this is another aspect that we're
looking at with respect to -- I mean, it's -- and be
aware that that's for 1 percent of the year at that
high number, so --
MR. LEVENSON: No, no, I was looking at
the 4.5 to 9, which is 20 percent.
MR. HILL: That's total suspended
particulate, so an appreciable portion of that doesn't
go into the lungs.
MR. McCARTIN: That's what I was trying to
get to, is the -- one of the things we want to look at
-- I know DOE accounts for a much lower value than we
would for what gets ingested, but some of this is
larger particles that get into the nose and ultimately
get ingested.
My understanding, in talking to our health
physicist, is that doesn't cause -- it still causes a
fairly significant dose, not as low as the DOE is
estimating. I'm not sure why those differences are
there, that's another aspect that we want to look at.
I don't know who has the more reasonable approach, but
we're looking at -- in my mind, I thought -- we looked
at particles that were inhaled through the nose and
ultimately ingested, not going into the deep lung,
caused about a factor of 2 lower dose than what was
inhaled into the lung. And so it isn't quite as
important.
For DOE, I thought I remember reading
something that was 3 orders of magnitude less, and I'm
not quite sure why that difference is there, and
that's something that I want to --
MR. LEVENSON: Well, that would be a big
function of time because if you are a couple of
thousand years down the road and most of what's left
is plutonium, it pretty much goes through your gut,
it's not absorbed at all, whereas in your lung you get
the dose. So that ratio would be a big swing with
time.
MR. McCARTIN: Yeah, and that's -- once
again, it's another thing we're looking at, but in
terms of deep ingestion into the lung, we're not
looking at. It's not causing a significant health
problem just because of the dust itself.
CHAIRMAN GARRICK: Well, thanks to us,
we've extended this a little, so maybe we'd better try
to wrap up this before we freeze to death.
Who controls the thermostat in this room?
Is there a reason that door is open?
VOICE: Yes, sir.
MR. LARSON: To let the heat in.
CHAIRMAN GARRICK: Oh, to let the heat in.
Oh, okay.
MR. BAHADUR: It is by design so that the
meeting will be really short.
(Laughter.)
CHAIRMAN GARRICK: It didn't work. Okay,
Tim.
MR. McCARTIN: I will try to go quickly.
The DOE approach is -- we're not that different, but
you can see we got to these numbers quite a big
differently. DOE has an outdoor value that they have
32 to 45 percent of the time, and an indoor value for
the remainder. However, they have a mass loading that
applies for 10 years. After 10 years, they're at the
background level, and so it drops off quite a bit
quicker than ours.
(Slide)
In terms of the risk insight, if I look at
the differences between our two approaches and be
aware that there still is a fairly fluid situation, in
my mind, because of the different approaches, also the
part that DOE puts all the radionuclides in the top
centimeter, there's a lot of little things there. But
I think, I'll say around a factor of 5 difference in
risk.
What are we going to do? we're going to
start -- continue to look at analog systems for what
is the right mass loading for different situations.
As I mentioned, Brit, they've done some measurements
at Sera Negro that can be used.
What's the evolution of the ash particle
over time? As was indicated, once you get up in the
air is a big part of this calculation, when this ash
is deposited. How much of this and how does it change
the particle size, that also is an important part.
See, there's a lot of subtleties here that need to be
considered.
And the last one that we're looking at is
there has to be a correlation for how long the deposit
persists and how dusty it is, and there certainly
should be some -- if there's a lot of it in the air,
the wind is blowing, it shouldn't last very long, and
we'll be looking at that aspect of it for
reasonableness, how long the deposit continues, and
that's where the remobilization and the inhalation --
this is one where there's just a lot of
interconnection between what assumptions are being
made.
And the next two slides is really -- and
I'll try to go through them very quickly, and I think
I can --
CHAIRMAN GARRICK: They are very good, and
they are quite self-explanatory.
(Slide)
MR. McCARTIN: Right. And what this --
this is sort of a preview for the future, and what
we're trying to work on is a simple explanation of the
approaches in the TPA code, where the uncertainties,
how we got to where we are. I think we would like to
do this for both our own code, but also for the DOEs,
possibly a way to provide a very simple, quick way of
explaining, understanding of the Repository. And
rather than going through the igneous slide, which
really is in some areas pretty much a summary of what
you just heard, I will skip past that and go --
(Slide)
-- one of the things we're hoping to learn
from this exercise is that, indeed, as Marty
indicated, we're treating uncertainty in a similar
fashion. That may not be the case everywhere, but I
think it is important to understand how we're dealing
with uncertainty. And this may be -- I said earlier,
when we're a little further along with this for the
entire TPA code, it may be useful to come back to the
committee and go through this and get feedback.
There's a lot of different areas of the TPA code.
For example, what really matters for waste
package? Well, if you have a waste package that you
think is going to last past 10,000 years, initial
defective packages are quite important. the question
is, the uncertainty in estimating that is the
manufacturing process, the closure, weld defects, are
things you have to consider. Right now, in the TPA
code we have -- we're going from 0.01 percent to a 1
percent failure based primarily looking at some analog
manufacturing information.
CHAIRMAN GARRICK: Boy, that's a big jump.
MR. McCARTIN: Yes. Well, on average,
0.01 percent.
In terms of the corrosion, as we
indicated, what are the big uncertainties there, and
certainly the hydro-chemical/thermal environment is
the key there, and certainly we have the short-term
measurements, and the effect of the welding, and the
post-welding, et cetera, what's the going to do to the
lifetime.
Our current approach is that the corrosion
rate is based on the relative humidity, however, how
do we account for more aggressive water chemistries?
We have a chloride multiplication factor that
increases the chloride content and thereby affects the
waste package failure time. Right now, failures, as
indicated here for uniform passive corrosion, it's on
the order of 10,000 to 45,000 years.
Dripping on the waste package. Where will
packages get wet? You have possibilities in the
hydrologic environment for focusing flow. You also
have diversion, capillary diversion of water. And so
you've got a couple of processes there. In terms of
the fraction of waste packages that get wet, we have
a very broad range there. It's all of them or none of
them. We sample between the two. There is a lot of
uncertainty three, and you can see we've taken the
full range.
However, within that, there is -- I'll
maintain it's Dick Codell's finest moment at NRC. He
may not think so, but in terms of how do you try to
represent this dripping aspect, and he came up with
some factors that account for how much water is
dripping onto the waste package, and it's correlated.
Although we do sample between zero and 1, when you get
close to 1 where all the packages are being dripped
on, it's a little bit of water. Clearly, you have
drips everywhere. You have a finite amount of
infiltration. The amount of water than can drip on a
package is small.
On the other end of the spectrum, when you
get to very few waste packages, you have some focusing
of flow, so you get a lot of potential for more water
dripping on the containers because you have very few
getting wet, but obviously where it's flowing there
has to be more water to account for -- and I'll turn
to Bill's infamous conservation of mass. We've got to
account for the water. You can't have a lot of water
on a lot of packages. It's a lot of water on a few
packages or a little bit of water on a lot of
packages, and that really is what that tries to
account for.
Right now, how does that vary? We go from
basically 1 percent of the infiltration rate when a
lot of them are getting wet, to 3 times when very few
packages get wet. So, you can see the range of the --
MR. LEVENSON: Three times as much water
reaches the package and comes in?
MR. McCARTIN: No, 3 times the
infiltration rate.
MR. LEVENSON: It would be more than
filtration rate.
MR. McCARTIN: Not in volume, it's focus.
MR. LEVENSON: Oh.
MR. McCARTIN: The infiltration rate is --
CHAIRMAN GARRICK: Funneled.
MR. McCARTIN: Right. In terms of the
drip shield, right now in our code, we have no
mechanistic model for its failure. We specify a
failure of time, which is approximately 5,000 years.
CHAIRMAN GARRICK: That's quite a bit
shorter than TSPA/SR at least.
MR. McCARTIN: Yeah, I think so.
CHAIRMAN GARRICK: It was about 20,000
years, as I recall.
MR. McCARTIN: But you can see, for us --
once again, getting back to what causes corrosion,
it's the humidity. So, whether we have drips or not
really doesn't affect the corrosion rate as much, but
in terms of can spent fuel, if there is an initial
defect, can water get in to mobilize the waste, the
answer would be no. And as I indicated, we do not
have a diffusive release, we only have an invective
release, so you do have to have dripping water. We
sort of felt without dripping water, you weren't going
to get anything out of the waste package. DOE would
have releases from drips.
CHAIRMAN GARRICK: So the drip shield
serves your model.
MR. McCARTIN: Yes.
CHAIRMAN GARRICK: And essentially does
not serve the DOE model.
MR. McCARTIN: No. I mean, if they had a
defective canister underneath the drip shield, even if
the drip shield was intact, the way I understand it,
they would still have releases.
CHAIRMAN GARRICK: Well, they are based on
diffusivity transport.
MR. McCARTIN: Yes. Spent fuel cladding,
as I indicated, right now there is the unzipping of
the clad. We think it's fairly uncertain what kind of
zipping would occur over hundreds of years, and just
thermally you're not going to see dripping for quite
a while, so you're looking at this not really as put
into the repository, but quite a few hundred years
afterwards. We right now take no credit for --
CHAIRMAN GARRICK: And that's kind of
evidence supported. There are thousands of assembly
years of experience with no such phenomena observed.
MR. McCARTIN: And then the final one is
diffusion of the radionuclides from the waste package.
There clearly is a very complex chemistry inside the
waste package, however, when we've done our analyses,
when you have dripping into the waste package, the
invection will dominate.
We have not included diffusion when -- you
get a hole and all of a sudden things start diffusing
out. we just don't include that, but we're aware of
it. We may do some additional analyses along these
lines because of the DOE model. At one time, be aware
that we did have it in our TPA code. We removed it.
It cost a lot of computer time and produced such a
small release, we said why bother. We may add it back
in, and we'll look at it, but --
CHAIRMAN GARRICK: This was the bathtub
model.
MR. McCARTIN: No. I mean -- well, we saw
the bathtub model, but a diffusional release out of
the waste package at one time we had. We removed it.
We were thinking about putting it back in due to the
DOE model, but I think the important thing is we
recognize the differences. DOE appropriately shows us
what the diffusional release is, the invection
release, so we can understand the differences, but
it's an interesting aspect between the two.
With that, like I said, we hope to do that
for all the codes, and this was really a very quick
cut. We've just begun it. We hope to improve the
detail and information on there.
(Slide)
With that, let me wrap up very quickly.
What we're trying to do is improve our understanding
in areas important to representing the consequences
from magmatic events and --
MR. LARSON: Time, as a result of that
improved understanding, have you come up with any FEPS
that the DOE has disregarded that should not be
disregarded? Is this being looked at?
MR. McCARTIN: Well, we certainly look at
the FEPs. In a general sense, I'd say no. They
certainly are looking at the alternative flow paths as
another -- although, there's a part -- is that a
different FEP, or is it part of igneous activity?
And I'm really not aware of any
significant FEPS that we've seen. And how things will
change, if I had to give my gut reaction as we go
through all this work, I think it's possible one could
see a 1 to 2 order of magnitude change. Time will
tell. Like I said, there's a fair amount of work we
hope to accomplish this year.
I'll make the offer -- none of my
management are here, so I can speak freely. I think
it would be fair, as this work improves, we can come
back and report on this, and certainly I think Brit
and John Trapp, as the aeromag data comes in, that
clearly mobility is not my strength by any means in
terms of how those might change. It might make sense
when that is, as that's analyzed, to come back also.
CHAIRMAN GARRICK: A couple of closing
questions. How many agreements are outstanding on
this KTI?
MR. HILL: We had 22 to begin with. I
believe 6 are completed -- 7 are completed -- close
enough.
CHAIRMAN GARRICK: If all of this goes in
the wrong direction, are we heading in a direction
that could make this a possible problem?
MR. McCARTIN: Well, I guess my impression
of the estimates are I find it very hard to believe
that these things at all go in one direction. I think
that is a very -- although there's no guarantees in
life, but I think it would be very, very odd that they
would all go in one direction or the other. And so my
personal take is that I think we're probably closer to
maybe it stays the same and maybe it drops an order of
magnitude.
MR. LARSON: So your estimate there is not
net, but individual estimates?
MR. McCARTIN: No, it's net. If I had to
be pinned down, and that's what this is trying to do,
I'd say 1 to 2 orders of magnitude, and I think that's
the net change. Which direction is it? I think most
likely it's the lower, my gut reaction at this time.
There are things that could make it higher. You saw
a combination of both. Probability could increase.
Number of waste packages might increase. I think mass
loading probably will be reduced. So, where you end
up I don't quite know.
CHAIRMAN GARRICK: Andy, are they wanting
anything from us on this, or is this informational?
MR. CAMPBELL: Well, this is primarily
informational preparation for future ACW reports to
the Commission, but if the committee feels that it
would like to weigh in on this issue, this is probably
the time to do that because this is a particular KTI
that's been talked about.
MR. LEVENSON: Can I ask one more question
out of ignorance, which is where most of mine come
from. Why is Mount St. Helen's not referred to as
anything that might be an analog, is it so different?
Probably better measurements on that than most.
MR. McCARTIN: Well, in terms of -- Brit
probably can answer this better than I, but I'll put
my two cents in first and then go to Brit.
I do know DOE, in their reports, they
certainly cite Mount St. Helen's as the basis for
their ten-year -- after ten years reducing the mass
loading to the background. And so they are looking at
some of that information. There are aspects of the
eruptiveness that are -- in terms of the mass loading,
there are some limitations of Mount St. Helen's.
If you go near there in terms of --
there's a lot of education, different soil types than
Yucca Mountain, et cetera, that you have to take into
account when you look at if the ash can be easily
incorporated into the soil there versus possibly at
Yucca Mountain soils, and that's something that has to
be considered. But I think we try to look at
everything that's analogous, and I'll ask Brit if he
has some other thoughts.
MR. HILL: Just a couple of very quick,
off-the-top points. First, the grain size character
of the Mount St. Helen ash, especially in the areas
where the occupational studies were done for particle
concentration. It's fundamentally very different from
what we see from basaltic ash.
Second, it is highly dedicated, well
established soil that receives a lot more rainfall
than we see from the Yucca Mountain sorts of
scenarios. So the granulometry is different, the
depositional area is different. The volume away from
about the initial deposit itself is fairly small and
very thin deposits. On slopes and terrain it really
was very stable to begin with, and also very permeable
to begin with, that allow a lot of infiltration to
occur.
So, while we can gain some general
insights from the Mount St. Helen's information,
trying to understand what's happening between zero and
20 kilometers away from the Yucca Mountain system has
some fairly significant limitations.
MR. LEVENSON: We're assuming no climate
changes for these 10,000 years, right?
MR. HILL: We're looking, though, at these
peak risks from volcanic disruption occurring on the
order of the first thousand years. So, while climate
change would affect the 10,000 year view for the
period that we're most concerned about, we're not too
worried about climate change. There's at least no
evidence to support significant climate change in the
first thousand years.
MR. LEVENSON: And that affects the
probability by a factor of 10.
MR. HILL: The probability is the same at
any given year. There's really no -- with the
homogenous fall recurring rate, there's no effective
difference between the probability in year 100 and the
probability in year 10,000.
MR. McCARTIN: When we do our
calculations, we're looking at year one, the
probability. We're not, say, taking the event,
multiplying by 10,000 years, and getting a single
probability to then say -- and I don't know, maybe Dr.
-- this may not be as clear. We don't say take 10-7
times 10,000 so the probability of that event --
CHAIRMAN GARRICK: No, I understand.
MR. McCARTIN: -- okay -- at year one,
it's 10-7.
CHAIRMAN GARRICK: Right. Okay. Any
other questions from any of the members, consultants,
staff? Yes?
MR. HAMDAN: Tim, just one question. I'm
not clear on the one aspect of this. Are we saying
that every igneous activity will have an intrusive
component and risk, and extrusive also?
MR. McCARTIN: Yes.
MR. HAMDAN: So, how are you deciding how
much of the event is extrusive and how much is
intrusive? Do you assign probabilities, or how do you
go about doing that?
MR. McCARTIN: For us, right now, we
merely -- we have -- we specified a number of waste
packages that will be affected by intrusion. We have
not worried that much about assigning a specific
probability, we just -- we can allow it to occur at
the same time we do the eruptive one, but it tends to
be -- relative to the eruptive one, it's a very small
risk value.
MR. HAMDAN: So, how about an intrusive
event -- how about an igneous activity when that does
not include eruption?
MR. McCARTIN: It could be a factor of 10
or higher in terms of probability.
MR. HAMDAN: Is that modeled --
MR. McCARTIN: We can. I mean, right now
the intrusive event is for us a very small amount.
DOE obviously has a larger one and, like I said, we'll
be looking at those differences. It's quite possible
we have it wrong. If it comes up in consequence, then
maybe we need to look at it a little more carefully.
But, right now, we see it dominated by the eruptive.
CHAIRMAN GARRICK: Andy?
MR. CAMPBELL: I was just going to ask if
there was anything to be gained in terms of ash
distributions from looking at similar type of basaltic
cones, signatory cones, in the basin and range. You
mentioned Sunset Crater. Is there anything to be
gained from studying the volcanoes that are there at
Yucca Mountain in Crater Flats? Are they so old that
anything that you can interpolate to 10,000 years has
long since gone?
MR. HILL: Right. The youngest volcano is
80,000 years old, and almost all the ash deposit is
gone, there's just a fragment here and there. In
terms of the basin and range, the Sunset Crater is
about 1,000 years old. It's much larger volume. But,
again, one of the two volcanoes of similar type that
still has a preserved deposit, we're not using it as
a direct analogy, we're trying to look at how tetra
would behave, but not how the eruption would progress.
That's why we've been using the historical
analogs, because we have nothing in the basin and
range that's historical, and most of the deposits are
before the highly eroded condition, so we look at much
younger deposits on the order of 50 to erupting while
we watch, so that we can better understand the
dispersive processes.
But looking at places like in Mexico or
Sunset Crater and then comparing those to the older
volcanoes in the base and range, that's where we came
up with the scaling for at 1,000 years there seems to
be some deposit left, at 10,000 years it seems to be
pretty well gone. So, the 1,000 year half-life that
we're using is really a scale of 100 year, 1,000 year,
10,000 years. You see it at 1,000, we don't see it at
10-, and clearly it's there at 100. So, 10,000 year
it's all gone, given a half-life of 1,000, that seems
to be the best first pass.
MR. CAMPBELL: One other question for Tim,
on these tables that you developed at the end of your
presentation, is there any thought to developing a
column to identify built-in conservatisms, and this is
an issue the committee has been dealing with, and kind
of raised it in the context of TSPA/SR of identifying
conservatism. Has the staff thought about it, you
guys thought about, you know, looking at how
conservatisms are being treated and carried through
the model? Is that part of all this?
MR. McCARTIN: We certainly will look into
that. I'm always trying to think of a better word
rather than "conservatism". I don't think there is
one, unfortunately. The only reason I say that, we
need to get to that, but part of it is the information
you have, and some of it is you're making some
assumptions and, as Dr. Garrick said, assumption base
versus evidence, so often --
CHAIRMAN GARRICK: What might be better is
a cryptic identification of the relevant evidence.
MR. McCARTIN: Yes. Absolutely.
CHAIRMAN GARRICK: Supporting evidence.
MR. McCARTIN: And that is one there that
here's what the evidence tells us, and let someone
draw their own view of whether it's conservative or
not.
CHAIRMAN GARRICK: Any other questions
from anybody?
(No response.)
Okay. I want to thank Brit and Tim for a
very informative hour presentation that went a couple
of hours, and we look forward to following up with all
the path forwards that you identify, and we also want
to thank our consultant for his significant
contribution in this session, and we will now take a
recess and when we come back -- this terminates all of
our briefings and discussions of this type.
We will now go into our usual letter
discussion and letter-writing session, for which we
will not need reporting, and until then we will take
a 15-minute break.
(Whereupon, at 3:40 p.m., the briefing
session was concluded.)
Page Last Reviewed/Updated Monday, October 02, 2017